Correction des données et preuves

Publish six reviewed QS35–40 V6 corrections with explicit held/history, source scope and confidence limitations.

Le résumé est affiché en anglais.

Correction historique : les nombres et affirmations correspondent à cette version, pas à la version actuelle.

2026-10-22T13:36:57.019481Z · qs35-40-v6-r3-20261022

• McGill University · 12 reviewed records in this revision · Voir la page de politique
  • Enregistrements ajoutés: 0; Enregistrements corrigés: 0; En attente / exclues des conclusions actuelles: 0

Historique de la version V6 précédente

Record revision, scope and original evidence

Original evidence and audience, course and unit scopes are preserved. Tracker correction time is not a policy effective date.

The records below reflect this historical version, not the current policy record or current review conclusions.

73eec78c37f29d8f55abd2f7408746baecb31347ad105a541ae559d11a218753

{
  "addedClaimIds": [],
  "updatedClaimIds": [],
  "heldClaimIds": []
}

• mcgill-claim-approved-tools — McGill IT Services offers a secure Commercial Data Protection version of Microsoft Copilot, accessed with McGill credentials, which it chose because it conforms to the University's privacy and data protection standards; the public version of Copilot does not conform and should not be used for McGill-related purposes. Even in the secure version, users should not share private, sensitive, PII, PHI, PCI or other confidential information. McGill's separate 'AI Tools at McGill' directory returned HTTP 403 when re-checked on 2026-10-20, so the current directory entries (names, audiences, prices) could not be independently re-verified. Tool approval is not permission to use AI in an assessment.
• mcgill-claim-privacy — McGill guidance says users should mitigate potential privacy concerns by removing personally identifying information when using AI tools, be careful with sensitive or restricted material, and avoid using Personal Health Information (PHI) or Payment Card Industry (PCI) data with AI tools.
• cl-mcgill-university-microsoft-copilot-184 — McGill IT Services' June 2024 article describes the secure Commercial Data Protection version of Microsoft Copilot, accessed via Microsoft Edge, as having undergone a data security and risk assessment and as the version McGill chose for McGill-related work. 'Commercial Data Protection' is the product name used in that 2024 article; the current Microsoft product name and McGill directory entry were not re-verified because the directory returned HTTP 403.
• mcgill-claim-academic-integrity — McGill's Provost-endorsed principles state that instructors remain responsible for comporting themselves according to the highest standards of academic integrity in their use of generative AI tools. Instructors must be explicit in course outlines about the expectations for use of generative AI tools and may set limits on their use in assessment tasks.
• mcgill-claim-teaching — McGill recommends that instructors explain to students in their course outline what the appropriate use or non-use is of generative AI tools in the context of that course. The use or non-use of these tools should align with the learning outcomes associated with the course.
• draft-claim-mcgill-code-ai-offences — McGill's Code of Student Conduct and Disciplinary Procedures (current version effective November 12, 2025) states that no student may use generative AI in an assessment in a manner not authorized by the relevant instructor or supervisor, or that violates the course's AI requirements (Article 25(b)), and that using any tool, including generative AI, that has been disallowed in the course outline or written assessment instructions is cheating (Article 26(f)). AI-detection software is not sufficient by itself to substantiate an allegation under Article 25. Permission is therefore set by the instructor, supervisor and course, not university-wide.
• draft-claim-mcgill-code-procedure-appeal — Under McGill's Code, no allegation is substantiated unless the University proves it on a Balance of Probabilities, and the student is entitled to have an Advisor present at any interview or hearing under the Code. A party wishing to appeal a decision of the CSD Hearing Committee must notify the Secretary of the Appeal Committee within 10 Days (working days under the Code) of receiving official notice of that decision; this deadline applies to CSD Hearing Committee decisions, not to every academic decision. Appeals are limited to new evidence, procedural error or bias, or fundamental error in interpreting offences or applying sanctions.
• draft-claim-mcgill-student-responsibility-principle — McGill's Principles on Generative AI in Teaching and Learning state (fifth principle) that students remain responsible for academic rigour: verifying the accuracy of AI-generated information, acknowledging the use of generative AI tools where applicable, informing themselves about and complying with instructors' explicit expectations, and respecting limits set on AI use in assessment tasks. Whether AI is allowed or prohibited is not set university-wide.
• draft-claim-mcgill-ai-detection-not-supported — McGill teaching guidance (August 2026) states that McGill does not support the use of generative AI detection tools to determine whether student work was generated using AI, because they can produce unreliable results and should not be relied on as evidence of unauthorized use. This teaching position complements, but is distinct from, the Code rule that detector output alone cannot substantiate an Article 25 allegation.
• draft-claim-mcgill-gps-research-risk-tiers — McGill Graduate and Postdoctoral Studies research guidelines (November 2024) sort graduate students' AI uses into three tiers: low-to-moderate risk (e.g. proofreading, light editing, debugging existing code, formatting references), moderate-to-high risk (generating content for one's own use, such as summaries for personal study), and very high risk (generating content and presenting it as one's own work, which constitutes plagiarism under the Code of Student Conduct, e.g. writing entire sections). These guidelines are research-focused, not course-based.
• draft-claim-mcgill-thesis-disclosure-recommended — Since May 1, 2026, McGill GPS asks thesis students to consult its GenAI Research Statement and Disclosure Guidelines and, with their supervisor, include a GenAI statement and disclosure framework in master's and doctoral theses. This statement and framework are recommended, not yet mandatory; failure to disclose AI use may compromise or delay assessment of the thesis, and after Fall 2026 consultations they may become mandatory. Department, supervisor and discipline requirements still apply.
• draft-claim-mcgill-tool-review-translation — McGill IT Services states that, unless formally reviewed and approved by McGill, it is prohibited to use any generative AI tool for McGill-related purposes; new AI tools, AI features added to existing platforms and requests involving sensitive data go through the AI Review Committee process. For translating McGill content, Linguistic Services lists two approved tools - DeepL Pro (paid version) and Microsoft Copilot (enterprise version) - with the free versions prohibited for data-security reasons and no private or sensitive information to be entered. These are institutional data and translation rules, not course permissions.